There are families of sensors such as gyros and accelerometers which require the angle measurement of a rotating member for their output. A schematic representation of a typical rotating member 10 is shown in FIG. 1 in a differential arrangement of capacitive plates 11.
Capacitive sensors are used because of their low noise characteristics, size, stable dimensions, and potentially very small gaps. Noise levels on the order of 10 nano-radians at a bandwidth of 100 hertz are possible.
The major limiting factor in the performance of capacitive pick-offs is the effects of stray capacitance pickup on the electronic circuitry which handles the rotation signal. This is particularly true when the capacitive values are on the order of a few pico-farads as they are in small gyros and accelerometers fabricated by micro- and milli-machining technologies. A solution is to couple the capacitive transducer 12 to a differential transformer 13. The capacitor pairs are tied in parallel with outputs connected to opposite ends of a transformer primary with a center tap to ground.
FIG. 2a illustrates the full tilt pick-off which includes the capacitive plates 12, the Differential Current Transformer (DCT) 13 and current to voltage electronics 14. FIG. 2b illustrates the full tilt pick-off which includes the capacitive plates 15, the planar DCT 16 and current to voltage electronics 17.
In the DCT, two currents flow to a common tap to ground. When the capacitive transducer is tilted, one of the currents becomes greater than the other and a current is induced in the secondary. The resultant current is converted to a voltage using an operational amplifier.
It is desired to locate the differential current transformer (DCT) next to the capacitive plates in order to take the current difference at the signal source and amplify it before capacitive noise pick-up in the leads can become a problem.
For miniature, essentially planar devices which are fabricated with an assembly of planar layers, it is desirable for the DCT to also have a planar form so that it can either be formed or be placed next to the transducer. At this time custom-wound ferrite cores are used. Their shape and size make it difficult to place it sufficiently close to the transducer. In addition, the core winding leads are susceptible to pick-up themselves.
The planar transformer approach will allow inductive components to be fabricated directly onto parts in many applications. The prospect of integration with IC chips and package structures may replace the current approach of pick and placement of wound cores onto parts, and would result in a much more cost effective approach.
The planar MilliDCT technology provides paths to the design and manufacture of a wide range of inductive components, (such as power transformers, isolation transformers, chokes, filters, mixers, etc.), which have smaller footprints, flatter profiles, lower weight, lower manufacturing cost, and greater potential for integration than is possible with existing, machine-wound inductive coils.
In addition to angular detection, translations can also be measured with a suitable differential configuration of capacitive plates.
Traditional Approach to Planar Transformers
Planar transformer designs using MEMS (Microelectromechanical Systems) technologies have been published. FIG. 3 shows a toroidal concept 26 which is representative of the state of the art. It consists of a magnetic core in the shape of a thin circular ring 22 sandwiched by polyimide electrical insulators 21, 23. To form a DCT, this assembly is encircled by three coils 20, 24, two primaries and one secondary (detail not shown) with the condition that the primaries are matched.
The fabrication approach involves the electroplating of the magnetic and conductor materials using patterning by photolithography. A monolithic approach with processes performed at each layer till the full device is formed.
Since the figure is of a blow-up of the concept, it is understood that the lower 24 and upper 20 conductors are connected so that the coils spiral around the core.
Parameter requirements are for: (1) near unity coupling between the primary and secondary coils, (2) coil inductance, (3) conductor resistances and (4) quality factor, Q.
Induction calculations assume that the flux flows across the full cross-sectional area of the core which is encircled by the winding turn. However, with increasing operational frequency, the rate of change of flux with time increases, generating a current in the magnetic material which in turn forms its own flux that tends to oppose the original. The currents are called eddy currents. Unlike the conductive windings which are wound outside the material, however, the eddy currents occur at all radii inside the volume of the material. The net effect is that the flux is the maximum at the center and decreases with distance from it. Therefore with increasing frequency, the flux at the center is cancelled and eventually can be eliminated from the bulk. With further increase in frequency a condition occurs where flux only flows within a skin-depth of the surface. To get around the problem, the magnetic core is laminated in thinner dimensions on the order of the skin depth. The number of windings also determine the inductance.
The difficulties with this approach are numerous:
1. Multi-layering of coils is impractical in comparison to ferrite core transformers; essentially limited to one layer in this planar design. PA0 2. The number of coil turns is limited by the aspect ratio of the fabrication process otherwise a short can occur between them. PA0 3. In order to connect each set of upper and lower conductor segments, the alignment of subsequent masks needs to be extremely precise. For each winding turn, 4 connections are required. PA0 4. The interface between each conductor connection needs to be clean in order to minimize the electrical resistance. PA0 5. The matching between the coils is affected by 2-4 of this list. PA0 6. The Q is affected by 1-4 of this list. PA0 7. The induction is affected by 1-3 of this list. PA0 8. Leakage flux occurs since the windings do not totally enclose the magnetic core.
All the difficulties indicate an expensive and difficult process. Very high aspect processes like LIGA are necessary. In addition the difficulties mentioned limit how small the DCT can be made.